Abstract BACKGROUND Glioblastoma (GBM) is a fatal tumor whose aggressiveness, heterogeneity, therapy resistance and poor blood-brain-barrier penetration hinder the amelioration of the standard-of-care. Included in the GBM mass are the tumor initiating cells (TICs), representing the driver of GBM growth and relapse in virtue of their stem-like traits and therapy-resistance. Being constantly exposed to environmental stress, including nutrients deficiency, hypoxia and therapeutic insults, all GBM cells -and TICs in particular- have to be highly adaptive in order to survive. Thus, their ability to cope with stress could be targeted to curtail TICs maintenance and the whole GBM aggressiveness. The key of TICs adaptation relies, among the others, on their epigenetic plasticity, hence encouraging epigenetic drugs testing. MATERIAL AND METHODS By exploiting patient-derived GBM TICs and orthotopic xenograft models, we tested the antitumorigenic features of a novel, selective, orally bioavailable and brain-penetrant Lysine-specific histone demethylase 1 inhibitor (LSD1i). We confirmed the specificity of its effects by LSD1 genetic targeting. A combination of RNA-seq, Chromatin Immunoprecipitation(ChIP)-seq, Mass Spectrometry and reverse genetic experiments unraveled LSD1 molecular players in GBM TICs. RESULTS We identified LSD1 as a druggable target in human GBM: LSD1i treatment, mirrored by LSD1 genetic targeting, impairs growth, viability, stem-like traits and in vivo tumorigenicity of GBM TICs. Mechanistically, LSD1 is crucial for the expression of the activating transcription factor 4 (ATF4), which coordinates the integrated stress response (ISR) to manage stressful stimuli as nutrient deprivation and endoplasmic reticulum stress. By mimicking these stress cues in vitro, we found that LSD1i triggers a delayed but unabated ATF4 translation which provokes an over-lasting ISR, eventually culminating in GBM TICs apoptosis. Lastly, LSD1 demethylase activity is dispensable for ATF4 induction. Rather, LSD1i exerts its anti-tumorigenic potential by interfering with LSD1 scaffolding function in GBM TICs. CONCLUSION LSD1-directed therapy is likely a promising strategy to hinder GBM. By sensitizing GBM TICs to stress, LSD1i endangers the GBM TICs pool. The effectiveness of LSD1i administration in different patient-derived GBM TICs and xenografts, regardless of their molecular profile, places a strong rationale toward the clinical translation of this approach for GBM management. FUNDING Italian association for Cancer Research (AIRC) and Italian Ministry of Health